- This topic has 65 replies, 17 voices, and was last updated 19 years, 9 months ago by
.
-
Posts
-
The draw can have a real physical effect on the outcome of a race.
The fact that an 8-y-o has not won this particular race previously has no bearing on this race, but the fact that 8-y-os are usually past their best or vulnerable to less-exposed and younger horses, especially in races targeted for such horses, may well do.
In contrast to ACR1 I do not personally know of a single successful punter who pays more than lip service to trends as they are presented by the likes of Craig Thake in the Post.
Quote: from thedarkknight on 11:42 am on Jan. 12, 2005[br]
Does that mean I don’t want to back an 8 year, even if every other factor is in his favour and he has otherwise opposable horses to beat?<br>
TDK – not at all. It depends on whether you think you have found a reliable trend or not.
You could look and find that only 10 8yo’s+ had taken part in the race and all had started at 16/1 or bigger. That’s not a trend that would put me off.
Or you could find that there were 50 8yo+ contestants, 0 winners and a dozen that got beaten at 8/1 or less. That’s a trend it would pay to heed.
Having said that, it’s just a trend – a tendency rather than an absolute. You might decide that your horse is not a typical 8yo. For instance, he might have had less runs than a typical 8yo, he might be handled by a trainer who has shown he can overcome trends or you might decide that there is an overriding factor that’s more important on that particular day such as really heavy ground.
Trends are an extremely useful analytical tool, particularly with regard to jumping. <br>ACR1 and tdk are right however in that it is often the application/interpretation of the trend analysis which is at fault rather than the data itself.<br>It is vital, as ACR1 says, to position any data (e.g. no. of 8-y-o winners of a race) within the context (i.e. how many runners of that age were there). <br>To use an example. If I were looking at, say, the likelihood of horses at a specific price range winning a particular race I would analyse the lastten years results by filtering horses within price bands (for argument’s sake <2/1,3,1 to 9/2, 5/1 to 9/1, 10/1 to 24/1 and 25/1 upwards). I would then examine the individual results of each grouping, focusing on the market’s pre-race assessement (i.e. the starting price) against the actual result (the actual percentage wins of the horses in that grouping). Certain races consistently favour short priced horses and this analysis can reveal those and can reveal the edge which has historically been gained by focusing on those horses.<br>Statistically, using trend analysis in this way is flawed however. This is because the sample from which we draw our data is based on the set of results which have already occured. The race we are about to bet on does not form part of the sample we have analysed and what we are doing in effect is using statistical evaluation of a historical sample and making a BIG assuimption that the future sample will closely reflect that historical data set. This is not always the case, as we know, and this is why such analysis is inexact and problematic. <br>As tdk points out individual horses (he uses the example of Rooster Booster) can make a mockery of the trends. The important thing about trends though is that they can generally be reasonably accurate. So, while we may lose out in specific instances we may gain an edge in the long run. Again, as pointed out by tdk, intelligent use of trends where one can spot races where the trend may be upset because of a particular horse or circumstance can help make them even more profitable, although I have found that the trend quite often overcomes my subjective evaluation and I have thrown the baby out with bathwater.<br>To conclude, trends are powerful tools but, as with all types of backfitting, are statistically flawed and should not be used to the exclusion of other analytical tools, teh most important, IMO, being assessment of a horses ability.
Quote: from Prufrock on 12:08 pm on Jan. 12, 2005[br]
In contrast to ACR1 I do not personally know of a single successful punter who pays more than lip service to trends as they are presented by the likes of Craig Thake in the Post.
Pru- that’s not what I said. I said that "every successful punter is an analyser of trends".
You have to be or else you’re operating randomly. Successful punters are successful because they recognise what has been successful for them in the past. They spot patterns or trends and try to repeat them.
Sorry, ACR1.
I did mean to add "but this could just be a question of definition".
All but one of the successful punters I know of are pretty systematic in their approach, but not one of them bothers much with "trends" as Craig Thake would define the word, rather than as you would apparently define the word.
The crucial difference is between factors that may affect the performance of the horses in question directly—and in this respect it is entirely proper to analyse the draw, the abilities of the horses, pedigrees etc etc—and those factors that don’t, such as the SP and age of previous winners.
Quote: from cormack15 on 12:17 pm on Jan. 12, 2005[br]Statistically, using trend analysis in this way is flawed however. This is because the sample from which we draw our data is based on the set of results which have already occured.
No. It’s flawed because the sample size is 10.
Trends are the very last thing you take into considderation imo. If you went by trends, you would have diregarded Essex chances on Sunday immediately.
This, 8yo dont do well in this event is crap.<br>Even if an 8yo wins the race this year, for example, it can still state next year that 8yos dont do well in the event.
Personally I have had great success using trends at the big meetings both over jumps and on the flat as part of a wider analysis. Please don’t base your opinions of trends-based analysis on Craig Thake, the man’s a complete muppet and has no idea of how to interpret the data- he is the reason trends are still an edge for some of us.<br>I have a real problem with trends based on starting prices as statistically few outsiders win so that runs of favoured horses in races are common, but it only takes one long-priced winner to put you well ahead so the fact that no winner of a certain race had started at bigger than 10/1 in the past decade wouldn’t put me off a longshot: Think about what you’re evaluating here- the public’s ability to price a winner in the past versus the future??? <br>I have a calendar of races during the year which I find to be very reliable as being won by a similar profile of horse most years, or at least often enough to be profitable. My 2 favourite races are both handicap hurdles which are invariably won by a 5 or 6 year old carrying less than 11st: one is the opening race on Grand National day. I surmise that these are highly competitive contests which young unexposed animals win. The trend allows the field to be reduced to 3 or 4 runners most years and has been hitting the winner 90% of the time reliably for at least 12 years, at 25/1 in 2003. Obviously publications such as Weatherbys jumps and flat guides will reduce profitability over time but at places like Cheltenham and Aintree history does repeat itself because of the way trainers campaign horses year after year.<br>One trend which has more written about it than any is less than 11st in the National: People get slagged for quoting it but it is highly reliable and I won’t be abandoning it until it is broken.<br>I am not for a minute saying that trends alone are a sure way to profit but that in the right hands they can give you a  high percentage chance that the winner will come from a much reduced field, which has to be a big help.
trend before welsh national
no horse in the last 10 year’s has won carrying more than 10-11
sorry make that 11 year’s
That still wouldn’t put me off backing a horse carrying a big weight if it appeared to have the ability and aptitude to overcome that impost……….such as Carvill’s Hill in 1991………..
I’d be more interested (well, only a bit more to tell the truth) to know how well horses with big weights have run in the race over the years, rather than whether they happened to have won or not.
Most of us on here have been down this road on many occasions and the subject always provokes many varied replies.<br>I have a vested interest and thats well known but I just ignore anything except trainer stats. I do not expect to make a profit every day but do ok. <br>To me it does not matter if the horse is fav and has the best draw/ jockey/ positive on the exchanges etc etc if the trainer has not had a winner at the track in question it would be a ‘lay’ for me Im afraid.
I upset ‘smithy’ a while ago when I mentioned on this site that Brendan Duke does well OVERALL with his nat hunt runners as a he has only been going 3 seasons but has made a level stake profit in every season just under 200 points. Up to last season.<br>There was a recent winner at 66/1 (well backed at better prices on betfair by the way) a Cheltenham so no doubt he will return another l.s.p.<br>G. Butler overall does very well at the Ling A/W Track <br>he had 1 loser in season 1998 but every season since then he has made a level stake profit a £100 on every one at s.p. would have made over £5,100.00 . <br>I know Pipe had a winner at the Ling a/w the other day because I always lay his favs there but the last winner there was in season 1992 in season 1992 until the end of last season all the other 34 horse had been beaten many start quite short prices.<br>Even following Mark Johnston a high profile trainer blindly at Ascot since 1995 would show a l.s.p in 9 out of 10 seasons in the losing one he lost one point. But a £100 on every one at s.p would have made £17,886.0 How good is that trend? and thats why I am upset Ascot is out of action the next few seasons:(
Glenn – you are partly correct. The sample size is important but statistically sound conclusions can in fact be extrapolated from small sample sizes. I would also point out that while the number races analysed may be 10 the actual sample size depends on the number of runners analysed and this can run into hundreds in some cases so, although you are corect in highlighting that sample size is a factor which may (or indeed may not) affect the soundness of the statistical analysis, the size of the sample does not fundamentally flaw the statistics.<br>Analysis of a ten year sample of a particular race or races will allow you to draw sound statistical conclusions on the data within that sample. It does not, however, enable you to draw significant statistical conclusions on future races. It can only give clues and the stronger the significance, statistically, of a particular trend within the ten year sample the stronger the likelihood of it recurring in future samples of a similar nature although, statistically, we are on weak ground as there are a plethora of factors which can ensure that the ‘next’ ten years are fundamentally different than the previous ten.
In summary – sample size does not automatically flaw the data and prevent significant statistical conclusions from being drawn from it.
- You must be logged in to reply to this topic.